14:10:51 #startmeeting #openstack-cyborg 14:10:52 Meeting started Wed Aug 22 14:10:51 2018 UTC and is due to finish in 60 minutes. The chair is zhipeng. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:10:53 #info Coco_gao 14:10:54 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:10:56 The meeting name has been set to '_openstack_cyborg' 14:11:02 #info Sundar 14:11:11 #info wangzhh 14:11:13 #info Howard 14:11:14 #info xinran 14:11:17 #info sum12 14:11:31 Hi guys 14:11:49 sorry for the late, there was a fire drill in my building 14:12:01 Maybe something wrong with zuul just now? 14:12:12 Hi Li. 14:12:19 Hi Li. 14:12:42 #topic Rocky finish up 14:12:44 Hi, Li, hope all is good 14:13:09 Coco_gao status update on the ovo patch ? 14:13:22 #link https://review.openstack.org/593627 14:14:25 I updated the code, but still not tested yet. 14:15:49 What is our objective in Rocky -- to commit specific patches, or to be able to bring up a VM with some specific use case? 14:17:26 Sundar, to bring up the VM, we really need to finalized the spec with nova-cyborg interaction 14:17:37 Hi, Sundar, I read your driver-agent api spec, just don't have time to reply, I am sorry. 14:17:57 #info shaohe_feng_robo 14:18:00 Li_Liu: Thanks, that's what I thought. 14:18:04 #info Li_Liu 14:18:05 Coco_gao: NP, understand 14:18:44 Coco_gao, I will review your patch today 14:20:28 I still need to finish some functions, I commit this patch on Monday just want everyone to know the OVO first. 14:21:08 In case you have some commits for OVO. 14:21:20 comments, I am sorry. 14:21:25 Sundar, some data structure in Coco's patch is different from your new spec. 14:21:26 will we need to merge this patch this release? 14:21:31 I also have a patch coming up within couple days. it involves calling Agent from Cyborg to do programming 14:21:46 Could u help to review it? 14:21:59 zhipeng, can we still merge changes to Rocky yet? 14:22:10 Li_Liu: I think we need to align on the os-acc flow for programming too. 14:22:29 yes we still can 14:22:37 projects are doing rc2 now 14:22:50 wanhzhh: Sure, I should review Coco's patch. I need to catch up 14:23:07 Sundar: Thx. 14:23:22 I am trying to make sure if we should finish nova/cyborg interaction in Rocky 14:23:54 Sundar, yes, i agree, at least I wanna get the whole flow with rpcapi ready so that os-acc can call it when it wants to. 14:23:58 Honestly, IMHO, we are making piecemeal changes without the big picture :) It is important to get things done of course but, we hope it is in the right direction that will take us forward with Nova and others 14:24:34 Sundar, I don't know well on traits, so I don't know my OVO can meet the placement requirements. 14:25:04 Li_Liu: We should first agree on the os-acc flow, right? 14:25:44 Coco_gao: NP, I can help you. Power architecture developers want to support custom traits from the driver, as I mentione din os-acc spec 14:25:51 *sorry, discovery spec 14:26:19 By now, we don't have DB to store traits on Cyborg side. 14:26:40 Sundar, not only flow but also how to handle different hypervisor, different platform, etc. 14:27:21 Sundar, but currently feature I am implementing does not involve os-acc at all, its the REST api to allow fpga programming 14:27:27 Coco_gao: at least for now, we can store custom traits as attributes 14:27:58 Li+Liu: So, there is no expectation that we should use the exact same flow for os-acc, right? 14:28:00 OK. 14:28:00 Coco_gao, yes, put them in the attributes list 14:28:02 wangzhh: Agreed 14:29:23 We need to resolve: 1.report data to placement 2. os-acc flow(include some other things) 3. nova intergration 14:30:08 Sundar, yes, you can say that. but I think we could reuse that piece of code in os-acc anyways 14:30:23 wangzhh, exactly 14:30:31 wangzhh yes 14:30:59 xinran: good at all 1, 2, 3 14:31:30 1.report data to placement -- Coco_gao 2. os-acc flow(include some other things) -- wangzhh 3. nova intergration -- xinran... are these assignments right? 14:32:09 Li_Liu: all can be xinran . 14:32:37 lol 14:32:42 Wow,. 14:32:43 I think so 14:32:48 😂 14:33:10 I am starting design a new API for nova intergration 14:33:34 for she already do these things, we can avoid duplicated work. 14:33:35 with lock and batch alloc support 14:33:37 wangzhh, can you help xinran when you have some time? 14:33:51 Of course. 14:33:56 are there any implementation specific reviews of os-acc already ? 14:34:13 xinran: great 14:34:32 shaohe_feng_robo, it's great that xinran is the expert on all of these :) 14:34:49 plz do remember that we develop and implement what we agreed for Rocky, at the moment 14:35:04 anything beyond that is for design discussion starting at the Denver PTG 14:35:14 Li_Liu: she's new api is for the intergration 14:35:20 sum12, Yes, but It's piecemeal. 14:35:31 so I am assuming there will a batch of patch reviews coming in from you right? xinran 14:35:37 do not start development on those design specs that are not merged 14:35:57 zhipeng, Agree. 14:36:00 but pleacement report depend on driver 14:36:08 I just trying to understand if all these should be done in Rocky or not? 14:36:30 not sure all drivers are ready. 14:36:32 wangzhh: couple you point me to a piece ? I was expecting it to be a separate project 14:36:45 Sundar: what's the status of the new driver? 14:37:07 yes placement report depend on driver, did coco already have some works on it ? 14:37:57 Li_Liu: for interaction, edleafe can help on nova side, 14:38:24 found it 14:39:01 Li_Liu: yes, is it urgent? 14:39:21 sum12, OK. :) 14:39:35 xinran, just do it on your own pace, but do you have a eta tho? 14:40:17 Nguyen Hai proposed openstack/python-cyborgclient master: Update reno for stable/rocky https://review.openstack.org/586830 14:40:19 Li_Liu: who report to placement? 14:40:22 what kind of data placement need? I just store all of the data. 14:40:29 agent or conductor? 14:40:39 shaohe_feng: The driver developer is apparently busy with other stuff. He has reached a stage where rpm and deb packages for OPAE libraries get packaged with the tree. I'll sync up with him and help in testing. 14:41:00 besides, the driver-specific traits is not support right now. 14:41:02 Nguyen Hai proposed openstack/python-cyborgclient master: import zuul job settings from project-config https://review.openstack.org/592840 14:41:11 Nguyen Hai proposed openstack/python-cyborgclient master: switch documentation job to new PTI https://review.openstack.org/592841 14:41:11 Sundar: good. thanks . 14:41:16 Nguyen Hai proposed openstack/python-cyborgclient master: add python 3.6 unit test job https://review.openstack.org/592842 14:41:21 Li_Liu: i am afraid that it's quite hard to implement all in this release ... 14:41:54 xinran: agreed 14:41:57 yeah, I think the work to heavy for xinran 14:42:01 too 14:42:22 shaohe_feng_robo, agent report to cyborg, then cyborg uses the placement client to report to placement, right? 14:42:24 since the time is limited before Rocky 14:42:33 Coco_gao: have you already do some work on placement report which related to your current work? 14:42:41 Li_Liu: OK. 14:43:20 Li_Liu: yes, cyborg use placement client to report resource 14:43:21 xinran, I know, as I said, just do it on your own pace. but could you give estimations on them? 14:43:22 I didn't report to placement by now. I 14:44:37 I just collected the driver data, store in db. 14:45:20 every card is a sub-provider 14:45:25 Coco, After diff. :) 14:45:50 Li_Liu: I will probably submit first version of new api allowing nova/cyborg interaction next week 14:46:29 Coco_gao, that should be good for now. next we will figure out during the reporting step, how to invoke placement client apis 14:46:31 Li_Liu: the placement report does not need many code 14:46:46 xinran, sounds good 14:46:47 just need to know the deloyable well. 14:47:07 shaohe_feng_, yup 14:47:11 to be clear, did we agreed to improve the current API, I mean, add a new API like DeployableActions to do alloc and dealloc? 14:47:56 xinran, I suggest to imporve them in S release. 14:47:59 xinran: you can introduce your new API briefly 14:48:47 Li_Liu, where can I store the acc_topology data? 14:49:10 wangzhh: yes that what i desired 14:49:17 shaohe_feng_: sure 14:50:52 Folks, I want to clarify that all commits now should be considered provisional. We may need to change them substantially to meet the use cases and Nova requirements 14:51:04 Coco_gao, the acc_topology is not to be saved tho. It helps you to structure the deployable relationships 14:51:39 wangzhh: do you plan to pull in any patches for the Rocky release for os-acc ? 14:52:19 Coco_gao, in case it's not clear to you, we can discuss it further once you feel better 14:52:26 Sundar: will you work on the image upload ? auto parser the bitstream files, and generate the glance metadata? 14:52:46 sum12: No. 14:53:49 I am trying to add a new API representing all actions of deployables e.g. allocation/deallocation. for example: def post(self, body) with body like {"FPGA":"vendor_id":"","type":""...,"GPU","vendor_id":""...}. And in this method, deployable_list will be call with filters. 14:54:54 shaohe_feng_, Before upload image, we need download first, in agent. Do you think so? 14:54:57 Shaohe: Sure. 14:56:03 To support batch alloc/dealloc, there will be a loop to call deployable_list(...,filters) that's my current solution. Does anyone of you have some better suggestion? 14:57:54 xinran, for batch alloc/dealloc, we should have new APIs, and I think it's better to have a spec. 14:58:18 We should disscuss how to handle failure. 14:58:45 wangzhh, you mean the looping case? 14:59:17 Li_Liu, No, for the new design. 14:59:59 LI_Liu, I just mess up,can we list the remaining work and the timeline for that? 15:00:16 there is no spec currently 15:00:49 Oh, Sorry. Misunderstand u. xinran. 15:01:21 Coco_gao, you mean the remaining work on the agent reporting? 15:01:41 Current implemetion 15:01:43 on cyborg rocky. 15:01:56 doesn't need spec. 15:01:59 I need to leave now for the next call. Thank you all. Bfn! 15:04:30 Coco_gao, for rocky, in my opinion, I hope we can merge the ovo stuff that you are currently working on. anything on top of that is a bonus :) 15:05:15 plus, xinran's new api and my programming api 15:05:36 wangzhh, what's left on your plate for Rocky? 15:05:48 there are 1. driver discovery and data structure. 2. placement report. 3. nova integration (new api) and what else ? 15:06:33 Agent diff and save, related to Coco's patch. 15:06:47 I thought the placement client report is finished didn't it? 15:07:00 wangzhh, sounds good 15:07:54 but I am afraid, Sundar's Spec have difference with my code. 15:08:02 Li_liu, you mean we have finished reporting to placement? 15:08:34 Sundar's spec mention that diff should be done in conductor. Did we have an agreement on it? 15:08:34 the api 15:08:36 Not yet. xinran said , she is working on it. 15:08:54 https://github.com/openstack/cyborg/blob/master/cyborg/services/client/report.py 15:09:06 xinran, not in current release. 15:09:17 these code from shaohe is the placement client right? 15:09:36 nope, zhipeng think it should be done on agent in this release 15:09:51 ok thanks 15:10:09 Li_Liu: yes, client 15:10:18 Yes, but we should intergrate them in agent. 15:10:33 I think so 15:10:37 yup, it's just the matter when to use the client 15:10:50 wangzhh: we do not need to download 15:11:28 I think it should be used after diff. 15:11:52 shaohe_feng_, so what if we want to program? 15:12:01 Coco_gao: yes, exactly 15:12:24 Shouldn't we download bitstream first? 15:13:40 no 15:13:43 Bye, I need to go bed. 15:13:58 Coco_gao: have a good dream. 15:14:01 bye, Coco_gao, take a good rest 15:14:51 shaohe_feng_, wangzhh, are you guys talking about the same step here? 15:16:40 Maybe we can hold a zoom meeting later? 15:16:47 we should not be in same page 15:16:59 we mya be not in same page 15:18:52 ya, it's too late now, let's wrap up. I think we are clearer on the work assignment for R. let's get it done 15:19:04 Yep. 15:19:34 xinran, let us know if you need help, especially wangzhh and me can off load some of your work, right wangzhh :) 15:19:55 Of course. :) 15:20:15 ok, thanks guys :) 15:20:49 one last thing. remember to post your topics for PTG no matter if you are coming to Denver or not 15:20:49 https://etherpad.openstack.org/p/cyborg-ptg-stein 15:21:44 and let us know if you are coming to PTG 15:22:05 Not sure. :) 15:23:50 ok, have a good night/day wherever you are :) 15:23:57 #endmeeting